Study And Application Of Personalized Retrieval Technology In Search Engine

Directing against that many current information retrieval systems improve their accuracy in search model and information processing course and that they don't pay users more attention, this thesis puts forward a new kind of intelligent information retrieval system, which is based on users personalization, and presents its theory model and detail structure. During the course of searching, It integrates users interests by using agent technology. So it can provide query result users need more precisely. Personalization is to attach user the primary important position and really "analyze, understand" users requirement according to their background. This system avoids the fact that when different user uses same query, the result isn't different and that when same user uses same query in different time, the result will be same generally.First the thesis analyses the developing situation, working mechanism of search engine and points out its evaluating criterion, studying point and weak point. Then it presents a personalized intelligent search engine and introduces its theory and technology, which are used to design search engine. Furthermore the thesis analyses the differentia between intelligent search engines and the old ones in retrieval technology and setting up index. So strongpoint of intelligent search engine is depicted obviously, Finally it probes into some implementing modes of personalized search engine using agent technology and Java is used to program and actualize this system.When programming, in order to show "personalization" of search engine, this system sets up users modeling to register and manage users information, presents users interests learning and users interests judging algorithm and users information physical storage structure. At users interface, users can correct their interests database and can appraise the search results to rectify their interests degree. So interaction between users and system has been enhanced. Natural language is used so those fussy retrieval rules are avoided. Agent technology is used to analyze user's interests, feedback user's appraisal and provide user some interested information voluntarily using the form of E-mail. Therefore correlation between users and retrieval results has been improved. This search engine can greatly improve the precision of searching and provide users more precise personalized results. It can make Internet information retrieval more convenient and accurate.